[apple/network_cmds.git] / unbound / libunbound / unbound.h

/*
 * unbound.h - unbound validating resolver public API
 *
 * Copyright (c) 2007, NLnet Labs. All rights reserved.
 *
 * This software is open source.
 * 
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 
 * Redistributions of source code must retain the above copyright notice,
 * this list of conditions and the following disclaimer.
 * 
 * Redistributions in binary form must reproduce the above copyright notice,
 * this list of conditions and the following disclaimer in the documentation
 * and/or other materials provided with the distribution.
 * 
 * Neither the name of the NLNET LABS nor the names of its contributors may
 * be used to endorse or promote products derived from this software without
 * specific prior written permission.
 * 
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
 * HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
 * TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
 * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
 * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
 * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

/**
 * \file
 *
 * This file contains functions to resolve DNS queries and 
 * validate the answers. Synchonously and asynchronously.
 *
 * Several ways to use this interface from an application wishing
 * to perform (validated) DNS lookups.
 *
 * All start with
 *      ctx = ub_ctx_create();
 *      err = ub_ctx_add_ta(ctx, "...");
 *      err = ub_ctx_add_ta(ctx, "...");
 *      ... some lookups
 *      ... call ub_ctx_delete(ctx); when you want to stop.
 *
 * Application not threaded. Blocking.
 *      int err = ub_resolve(ctx, "www.example.com", ...
 *      if(err) fprintf(stderr, "lookup error: %s\n", ub_strerror(err));
 *      ... use the answer
 *
 * Application not threaded. Non-blocking ('asynchronous').
 *      err = ub_resolve_async(ctx, "www.example.com", ... my_callback);
 *      ... application resumes processing ...
 *      ... and when either ub_poll(ctx) is true
 *      ... or when the file descriptor ub_fd(ctx) is readable,
 *      ... or whenever, the app calls ...
 *      ub_process(ctx);
 *      ... if no result is ready, the app resumes processing above,
 *      ... or process() calls my_callback() with results.
 *
 *      ... if the application has nothing more to do, wait for answer
 *      ub_wait(ctx); 
 *
 * Application threaded. Blocking.
 *      Blocking, same as above. The current thread does the work.
 *      Multiple threads can use the *same context*, each does work and uses
 *      shared cache data from the context.
 *
 * Application threaded. Non-blocking ('asynchronous').
 *      ... setup threaded-asynchronous config option
 *      err = ub_ctx_async(ctx, 1);
 *      ... same as async for non-threaded
 *      ... the callbacks are called in the thread that calls process(ctx)
 *
 * Openssl needs to have locking in place, and the application must set
 * it up, because a mere library cannot do this, use the calls
 * CRYPTO_set_id_callback and CRYPTO_set_locking_callback.
 *
 * If no threading is compiled in, the above async example uses fork(2) to
 * create a process to perform the work. The forked process exits when the 
 * calling process exits, or ctx_delete() is called.
 * Otherwise, for asynchronous with threading, a worker thread is created.
 *
 * The blocking calls use shared ctx-cache when threaded. Thus
 * ub_resolve() and ub_resolve_async() && ub_wait() are
 * not the same. The first makes the current thread do the work, setting
 * up buffers, etc, to perform the work (but using shared cache data).
 * The second calls another worker thread (or process) to perform the work.
 * And no buffers need to be set up, but a context-switch happens.
 */
#ifndef _UB_UNBOUND_H
#define _UB_UNBOUND_H

#ifdef __cplusplus
extern "C" {
#endif

/** the version of this header file */
#define UNBOUND_VERSION_MAJOR @UNBOUND_VERSION_MAJOR@
#define UNBOUND_VERSION_MINOR @UNBOUND_VERSION_MINOR@
#define UNBOUND_VERSION_MICRO @UNBOUND_VERSION_MICRO@

/**
 * The validation context is created to hold the resolver status,
 * validation keys and a small cache (containing messages, rrsets,
 * roundtrip times, trusted keys, lameness information).
 *
 * Its contents are internally defined.
 */
struct ub_ctx;

/**
 * The validation and resolution results.
 * Allocated by the resolver, and need to be freed by the application
 * with ub_resolve_free().
 */
struct ub_result {
        /** The original question, name text string. */
        char* qname;
        /** the type asked for */
        int qtype;
        /** the class asked for */
        int qclass;

        /** 
         * a list of network order DNS rdata items, terminated with a 
         * NULL pointer, so that data[0] is the first result entry,
         * data[1] the second, and the last entry is NULL. 
         * If there was no data, data[0] is NULL.
         */
        char** data;

        /** the length in bytes of the data items, len[i] for data[i] */
        int* len;

        /** 
         * canonical name for the result (the final cname). 
         * zero terminated string.
         * May be NULL if no canonical name exists.
         */
        char* canonname;

        /**
         * DNS RCODE for the result. May contain additional error code if
         * there was no data due to an error. 0 (NOERROR) if okay.
         */
        int rcode;

        /**
         * The DNS answer packet. Network formatted. Can contain DNSSEC types.
         */
        void* answer_packet;
        /** length of the answer packet in octets. */
        int answer_len;

        /**
         * If there is any data, this is true.
         * If false, there was no data (nxdomain may be true, rcode can be set).
         */
        int havedata;

        /** 
         * If there was no data, and the domain did not exist, this is true.
         * If it is false, and there was no data, then the domain name 
         * is purported to exist, but the requested data type is not available.
         */
        int nxdomain;

        /**
         * True, if the result is validated securely.
         * False, if validation failed or domain queried has no security info.
         *
         * It is possible to get a result with no data (havedata is false),
         * and secure is true. This means that the non-existance of the data
         * was cryptographically proven (with signatures).
         */
        int secure;

        /** 
         * If the result was not secure (secure==0), and this result is due 
         * to a security failure, bogus is true.
         * This means the data has been actively tampered with, signatures
         * failed, expected signatures were not present, timestamps on 
         * signatures were out of date and so on.
         *
         * If !secure and !bogus, this can happen if the data is not secure 
         * because security is disabled for that domain name. 
         * This means the data is from a domain where data is not signed.
         */
        int bogus;
        
        /**
         * If the result is bogus this contains a string (zero terminated)
         * that describes the failure.  There may be other errors as well
         * as the one described, the description may not be perfectly accurate.
         * Is NULL if the result is not bogus.
         */
        char* why_bogus;

        /**
         * TTL for the result, in seconds.  If the security is bogus, then
         * you also cannot trust this value.
         */
        int ttl;
};

/**
 * Callback for results of async queries.
 * The readable function definition looks like:
 * void my_callback(void* my_arg, int err, struct ub_result* result);
 * It is called with
 *      void* my_arg: your pointer to a (struct of) data of your choice, 
 *              or NULL.
 *      int err: if 0 all is OK, otherwise an error occured and no results
 *           are forthcoming.
 *      struct result: pointer to more detailed result structure.
 *              This structure is allocated on the heap and needs to be
 *              freed with ub_resolve_free(result);
 */
typedef void (*ub_callback_t)(void*, int, struct ub_result*);

/**
 * Create a resolving and validation context.
 * The information from /etc/resolv.conf and /etc/hosts is not utilised by
 * default. Use ub_ctx_resolvconf and ub_ctx_hosts to read them.
 * @return a new context. default initialisation.
 *      returns NULL on error.
 */
struct ub_ctx* ub_ctx_create(void);

/**
 * Destroy a validation context and free all its resources.
 * Outstanding async queries are killed and callbacks are not called for them.
 * @param ctx: context to delete.
 */
void ub_ctx_delete(struct ub_ctx* ctx);

/**
 * Set an option for the context.
 * @param ctx: context.
 * @param opt: option name from the unbound.conf config file format.
 *      (not all settings applicable). The name includes the trailing ':'
 *      for example ub_ctx_set_option(ctx, "logfile:", "mylog.txt");
 *      This is a power-users interface that lets you specify all sorts
 *      of options.
 *      For some specific options, such as adding trust anchors, special
 *      routines exist.
 * @param val: value of the option.
 * @return: 0 if OK, else error.
 */
int ub_ctx_set_option(struct ub_ctx* ctx, const char* opt, const char* val);

/**
 * Get an option from the context.
 * @param ctx: context.
 * @param opt: option name from the unbound.conf config file format.
 *      (not all settings applicable). The name excludes the trailing ':'
 *      for example ub_ctx_get_option(ctx, "logfile", &result);
 *      This is a power-users interface that lets you specify all sorts
 *      of options.
 * @param str: the string is malloced and returned here. NULL on error.
 *      The caller must free() the string.  In cases with multiple 
 *      entries (auto-trust-anchor-file), a newline delimited list is 
 *      returned in the string.
 * @return 0 if OK else an error code (malloc failure, syntax error).
 */
int ub_ctx_get_option(struct ub_ctx* ctx, const char* opt, char** str);

/**
 * setup configuration for the given context.
 * @param ctx: context.
 * @param fname: unbound config file (not all settings applicable).
 *      This is a power-users interface that lets you specify all sorts
 *      of options.
 *      For some specific options, such as adding trust anchors, special
 *      routines exist.
 * @return: 0 if OK, else error.
 */
int ub_ctx_config(struct ub_ctx* ctx, const char* fname);

/**
 * Set machine to forward DNS queries to, the caching resolver to use. 
 * IP4 or IP6 address. Forwards all DNS requests to that machine, which 
 * is expected to run a recursive resolver. If the proxy is not 
 * DNSSEC-capable, validation may fail. Can be called several times, in 
 * that case the addresses are used as backup servers.
 *
 * To read the list of nameservers from /etc/resolv.conf (from DHCP or so),
 * use the call ub_ctx_resolvconf.
 *
 * @param ctx: context.
 *      At this time it is only possible to set configuration before the
 *      first resolve is done.
 * @param addr: address, IP4 or IP6 in string format.
 *      If the addr is NULL, forwarding is disabled.
 * @return 0 if OK, else error.
 */
int ub_ctx_set_fwd(struct ub_ctx* ctx, const char* addr);

/**
 * Read list of nameservers to use from the filename given.
 * Usually "/etc/resolv.conf". Uses those nameservers as caching proxies.
 * If they do not support DNSSEC, validation may fail.
 *
 * Only nameservers are picked up, the searchdomain, ndots and other
 * settings from resolv.conf(5) are ignored.
 *
 * @param ctx: context.
 *      At this time it is only possible to set configuration before the
 *      first resolve is done.
 * @param fname: file name string. If NULL "/etc/resolv.conf" is used.
 * @return 0 if OK, else error.
 */
int ub_ctx_resolvconf(struct ub_ctx* ctx, const char* fname);

/**
 * Read list of hosts from the filename given.
 * Usually "/etc/hosts". 
 * These addresses are not flagged as DNSSEC secure when queried for.
 *
 * @param ctx: context.
 *      At this time it is only possible to set configuration before the
 *      first resolve is done.
 * @param fname: file name string. If NULL "/etc/hosts" is used.
 * @return 0 if OK, else error.
 */
int ub_ctx_hosts(struct ub_ctx* ctx, const char* fname);

/**
 * Add a trust anchor to the given context.
 * The trust anchor is a string, on one line, that holds a valid DNSKEY or
 * DS RR. 
 * @param ctx: context.
 *      At this time it is only possible to add trusted keys before the
 *      first resolve is done.
 * @param ta: string, with zone-format RR on one line.
 *      [domainname] [TTL optional] [type] [class optional] [rdata contents]
 * @return 0 if OK, else error.
 */
int ub_ctx_add_ta(struct ub_ctx* ctx, const char* ta);

/**
 * Add trust anchors to the given context.
 * Pass name of a file with DS and DNSKEY records (like from dig or drill).
 * @param ctx: context.
 *      At this time it is only possible to add trusted keys before the
 *      first resolve is done.
 * @param fname: filename of file with keyfile with trust anchors.
 * @return 0 if OK, else error.
 */
int ub_ctx_add_ta_file(struct ub_ctx* ctx, const char* fname);

/**
 * Add trust anchor to the given context that is tracked with RFC5011
 * automated trust anchor maintenance.  The file is written to when the
 * trust anchor is changed.
 * Pass the name of a file that was output from eg. unbound-anchor,
 * or you can start it by providing a trusted DNSKEY or DS record on one
 * line in the file.
 * @param ctx: context.
 *      At this time it is only possible to add trusted keys before the
 *      first resolve is done.
 * @param fname: filename of file with trust anchor.
 * @return 0 if OK, else error.
 */
int ub_ctx_add_ta_autr(struct ub_ctx* ctx, const char* fname);

/**
 * Add trust anchors to the given context.
 * Pass the name of a bind-style config file with trusted-keys{}.
 * @param ctx: context.
 *      At this time it is only possible to add trusted keys before the
 *      first resolve is done.
 * @param fname: filename of file with bind-style config entries with trust
 *      anchors.
 * @return 0 if OK, else error.
 */
int ub_ctx_trustedkeys(struct ub_ctx* ctx, const char* fname);

/**
 * Set debug output (and error output) to the specified stream.
 * Pass NULL to disable. Default is stderr.
 * @param ctx: context.
 * @param out: FILE* out file stream to log to.
 *      Type void* to avoid stdio dependency of this header file.
 * @return 0 if OK, else error.
 */
int ub_ctx_debugout(struct ub_ctx* ctx, void* out);

/**
 * Set debug verbosity for the context
 * Output is directed to stderr.
 * @param ctx: context.
 * @param d: debug level, 0 is off, 1 is very minimal, 2 is detailed, 
 *      and 3 is lots.
 * @return 0 if OK, else error.
 */
int ub_ctx_debuglevel(struct ub_ctx* ctx, int d);

/**
 * Set a context behaviour for asynchronous action.
 * @param ctx: context.
 * @param dothread: if true, enables threading and a call to resolve_async() 
 *      creates a thread to handle work in the background.
 *      If false, a process is forked to handle work in the background.
 *      Changes to this setting after async() calls have been made have 
 *      no effect (delete and re-create the context to change).
 * @return 0 if OK, else error.
 */
int ub_ctx_async(struct ub_ctx* ctx, int dothread);

/**
 * Poll a context to see if it has any new results
 * Do not poll in a loop, instead extract the fd below to poll for readiness,
 * and then check, or wait using the wait routine.
 * @param ctx: context.
 * @return: 0 if nothing to read, or nonzero if a result is available.
 *      If nonzero, call ctx_process() to do callbacks.
 */
int ub_poll(struct ub_ctx* ctx);

/**
 * Wait for a context to finish with results. Calls ub_process() after
 * the wait for you. After the wait, there are no more outstanding 
 * asynchronous queries.
 * @param ctx: context.
 * @return: 0 if OK, else error.
 */
int ub_wait(struct ub_ctx* ctx);

/**
 * Get file descriptor. Wait for it to become readable, at this point
 * answers are returned from the asynchronous validating resolver.
 * Then call the ub_process to continue processing.
 * This routine works immediately after context creation, the fd
 * does not change.
 * @param ctx: context.
 * @return: -1 on error, or file descriptor to use select(2) with.
 */
int ub_fd(struct ub_ctx* ctx);

/**
 * Call this routine to continue processing results from the validating
 * resolver (when the fd becomes readable).
 * Will perform necessary callbacks.
 * @param ctx: context
 * @return: 0 if OK, else error.
 */
int ub_process(struct ub_ctx* ctx);

/**
 * Perform resolution and validation of the target name.
 * @param ctx: context.
 *      The context is finalized, and can no longer accept config changes.
 * @param name: domain name in text format (a zero terminated text string).
 * @param rrtype: type of RR in host order, 1 is A (address).
 * @param rrclass: class of RR in host order, 1 is IN (for internet).
 * @param result: the result data is returned in a newly allocated result
 *      structure. May be NULL on return, return value is set to an error 
 *      in that case (out of memory).
 * @return 0 if OK, else error.
 */
int ub_resolve(struct ub_ctx* ctx, const char* name, int rrtype, 
        int rrclass, struct ub_result** result);

/**
 * Perform resolution and validation of the target name.
 * Asynchronous, after a while, the callback will be called with your
 * data and the result.
 * @param ctx: context.
 *      If no thread or process has been created yet to perform the
 *      work in the background, it is created now.
 *      The context is finalized, and can no longer accept config changes.
 * @param name: domain name in text format (a string).
 * @param rrtype: type of RR in host order, 1 is A.
 * @param rrclass: class of RR in host order, 1 is IN (for internet).
 * @param mydata: this data is your own data (you can pass NULL),
 *      and is passed on to the callback function.
 * @param callback: this is called on completion of the resolution.
 *      It is called as:
 *      void callback(void* mydata, int err, struct ub_result* result)
 *      with mydata: the same as passed here, you may pass NULL,
 *      with err: is 0 when a result has been found.
 *      with result: a newly allocated result structure.
 *              The result may be NULL, in that case err is set.
 *
 *      If an error happens during processing, your callback will be called
 *      with error set to a nonzero value (and result==NULL).
 * @param async_id: if you pass a non-NULL value, an identifier number is
 *      returned for the query as it is in progress. It can be used to 
 *      cancel the query.
 * @return 0 if OK, else error.
 */
int ub_resolve_async(struct ub_ctx* ctx, const char* name, int rrtype, 
        int rrclass, void* mydata, ub_callback_t callback, int* async_id);

/**
 * Cancel an async query in progress.
 * Its callback will not be called.
 *
 * @param ctx: context.
 * @param async_id: which query to cancel.
 * @return 0 if OK, else error.
 * This routine can return an error if the async_id passed does not exist
 * or has already been delivered. If another thread is processing results
 * at the same time, the result may be delivered at the same time and the
 * cancel fails with an error.  Also the cancel can fail due to a system
 * error, no memory or socket failures.
 */
int ub_cancel(struct ub_ctx* ctx, int async_id);

/**
 * Free storage associated with a result structure.
 * @param result: to free
 */
void ub_resolve_free(struct ub_result* result);

/** 
 * Convert error value to a human readable string.
 * @param err: error code from one of the libunbound functions.
 * @return pointer to constant text string, zero terminated.
 */
const char* ub_strerror(int err);

/**
 * Debug routine.  Print the local zone information to debug output.
 * @param ctx: context.  Is finalized by the routine.
 * @return 0 if OK, else error.
 */
int ub_ctx_print_local_zones(struct ub_ctx* ctx);

/**
 * Add a new zone with the zonetype to the local authority info of the 
 * library.
 * @param ctx: context.  Is finalized by the routine.
 * @param zone_name: name of the zone in text, "example.com"
 *      If it already exists, the type is updated.
 * @param zone_type: type of the zone (like for unbound.conf) in text.
 * @return 0 if OK, else error.
 */
int ub_ctx_zone_add(struct ub_ctx* ctx, const char *zone_name, 
        const char *zone_type);

/**
 * Remove zone from local authority info of the library.
 * @param ctx: context.  Is finalized by the routine.
 * @param zone_name: name of the zone in text, "example.com"
 *      If it does not exist, nothing happens.
 * @return 0 if OK, else error.
 */
int ub_ctx_zone_remove(struct ub_ctx* ctx, const char *zone_name);

/**
 * Add localdata to the library local authority info.
 * Similar to local-data config statement.
 * @param ctx: context.  Is finalized by the routine.
 * @param data: the resource record in text format, for example
 *      "www.example.com IN A 127.0.0.1"
 * @return 0 if OK, else error.
 */
int ub_ctx_data_add(struct ub_ctx* ctx, const char *data);

/**
 * Remove localdata from the library local authority info.
 * @param ctx: context.  Is finalized by the routine.
 * @param data: the name to delete all data from, like "www.example.com".
 * @return 0 if OK, else error.
 */
int ub_ctx_data_remove(struct ub_ctx* ctx, const char *data);

/**
 * Get a version string from the libunbound implementation.
 * @return a static constant string with the version number.
 */
const char* ub_version(void);

#ifdef __cplusplus
}
#endif

#endif /* _UB_UNBOUND_H */